Tuesday, 10 December 2019

Thoughts on the SI-19BG tube

My seller has agreed to supply me another tube, on the presumption that the first has failed. This may not be the case, and if I discover otherwise I will return the spare, of course if I cant make either work I will have to return both.

A few thoughts on possible issues, other than complete failure -

1. Some tubes dislike certain current levels. A higher anode resistor may be required to get this tube to work;
2. Perhaps 400V is not enough, or maybe too much? Although if too much I would have expected it to avalanche. Characterizing it on a variable HV supply might find a working operating voltage. Available data shows 360 to 440V, and I have heard mention of successful operation at 440V.
3. Capacitance. Although too much shunt capacitance can kill tubes, most have an inherent capacitance of about 5pF - this tube being so tiny is 1pF. Perhaps a little extra capacitance will give it a charge 'kick' to help it to ionize.

Ive now received the 100Ω precision presets to allow me to properly control the 555 timer based HV supply. I will modify the Cockcroft doubler on this with an extra stage, to allow me to reach higher voltages. This I need to do anyway, as I also have a MST-17 tube coming, that requires 1600V!

Monday, 9 December 2019

SI-19BG tube disappointment

Earlier today I connected the tiny SI-19BG tube to the test circuit. Unfortunately, not one single count was registered. 

I obtained this tube from eBay seller "2015_gurie", and am now in contact  with the seller concerning a replacement tube.

It might be that I've just been unlucky and received a failed unit, or it might be an error in operation due to the lack of good information on these tubes.

I will transfer it to my variable HV supply and probe the tube with the scope, to see if there is any activity at all.

I've just today ordered a new, modern scope as well!

Sunday, 8 December 2019

Ooh! Remembered I have a check source!

Before going to bed, I tried out the RC pulse stretcher on the breadboard.  With a 100nF capacitor, and various resistors from 2k2 up to 4k7, I can say it had no detrimental effectseffect on the circuit.  Whether it actually stretches the pulses will need the oscilloscope to verify. 

Then I suddenly remembered - Bob M1BBV once gave me a WW2 fob watch! A bit of a hunt later (also finding the replacement strap for my Fitbit) I had it in hand and powered up the Geiger circuit...

...oh my! This thing is HOT! At no closer than 2 inches the circuit was going crazy! With the dial glass against the tube, the rate was almost more than the LED could cope with!

So, that's a nice bit of Radium I have.

RC constant practical use

I've been pondering ways to 'clean up' the pulses generated by the Geiger tubes, and make them longer and more uniform. To do this I've been revisiting RC constant and their practical applications. 

For the Geiger tubes, and RC circuit can form a pulse stretcher. The 1st transistor in my interface switches on the 2nd transistor, which switches the indicator LED and clicker. By adding an RC circuit between the two transistors,  I can stretch the pulses out by approximately the RC time constant. Since the tubes dead-time is 190uS, choosing an RC constant for 200uS makes a good pulse size.

The capacitor I added to cure the flickering problem was 22nF, and is probably acting as an RC pulse stretcher anyway with the 1st transistors collector resistor. 

This got me thinking of another,  really useful RC circuit.  Many years ago I used to service pendant radio alarms made by Tunstall telecom. These used a momentary push button to trigger them, and a clever RC circuit to hold the power on only long enough to complete the transmission cycle. This clearly has many interesting uses. The circuit shown is, I think, how it works.

Friday, 6 December 2019

Annoyed - wrong TIG rod!

The long awaited package from eBay seller "mitools" arrived today - and it's the WRONG item!

It should of course be a 2% thoriated rod, colour coded red, and 1.2mm thick. What they have sent is a 2% CERIATED 2.1mm rod, colour coded grey. This is useless as a check source for the blindingly obvious reason of it not being radioactive!

So it's going back for exchange  and perhaps sometime soon I might finally have a check source!

I've ordered some MPSA18 transistors as well to try the base control method of regulating the HV circuit.  These have very high gain. 

The circuit diagram of my current Geiger counter is shown below. I also need to add an external pulse output to drive a counter or meter.

Flicker problem solved

Adding a small bypass capacitor across the supply, in this case a 22nF polyester that just happened to be loose on the bench, made no difference to the flickering LED problem (which I discovered also manifest as a low rapid clicking from the loudspeaker!). So, I thought to try it instead as bypass across the G-M tube cathode resistor. 

This solved the flickering problem, but created another - no more pulses! 

Clearly this was absorbing the pulses without triggering the indicator circuit.  So I moved it instead to bypass the base of the 2nd transistor - bingo!

Good, solid flashes and clicks with each pulse, but no more dim flickering!

22nF might be too high a value, so at a later time I will try and find the lowest value that gives good results. 

Improved Clicker - and an Oddity...

The original single transistor driving the indicator LED and sounder was poor. I have improved this by using the first transistor to drive a second, which then switches the LED and sounder. I have also replaced the small sounder with a 1.5" cone 8Ω loudspeaker. The result is a much louder click and a nice bright flash from the LED, to indicate a detected particle/quanta.


I did attempt to use a trick I've seen on others builds of similar architecture, that is, using an extra diode and capacitor to obtain an extra supply from the HV circuits, higher than the supply but low enough for safe use for the indicator circuits. This gave me around 6V, but the current went through the roof (35mA) and it was clear that the drive oscillator was not cutting back as it should. For the moment I have abandoned this approach. The 3V circuit I'm using seems quite adequate.

But there is an oddity... the LED is always very dimly lit, and on close observation, can be seen to be flickering. Now, if I touch ANY part of the low voltage side of the system - the LED goes out, except when pulsed by the transistor switch. And I do mean anywhere! I can put a finger on the +Ve supply connection, or the ground, or the battery itself! However, if I touch the potential divider input to the 1st indicator transistor (i.e. the G-M tube cathode connection), the LED lights with a very bright fast flicker!

It seems quite obvious that the cause of this is leakage of the HV generators oscillator. I suspect that some bypass capacitors will be required to tame this.

In this configuration on the breadboard, the circuit draws just 1.63mA from the 3V supply. This is with the regulator transistor controlling the oscillator transistors collector. I might be able to get it lower, and improve the regulation, by controlling the base, but that's an experiment for later, as is lengthening the output pulses to a fixed pulse length for an external counter interface.

1st working Geiger counter mock-up

This is the first breadboard mock-up  Geiger counter working. Background count is about 20cpm using a BOI-33 G-M tube. Anode voltage 400V, supply 3V.

BOI-33 G-M tube mock-up
The indicator section of the circuit is a simple transistor switch. Its nowhere near good enough for a finished project - the LED flash is too dim and the sounder 'click' is far too quiet. There is also too much leakage through it meaning the LED is very dimly lit all the time. This section then needs more gain, and perhaps a pulse extender to allow a much more robust indication.

The current drawn from 2xAA cells in the absence of a pulse, is about 1.8mA. Its a bit higher than I would like, but quite acceptable. Ideally, I would get the quiescent current down to under 1mA, which might be achievable with some component value changes. Ive also tested this circuit with my huge STS-6 G-M tube, and it works great. Background count is appreciably higher of course, due to the much larger sensitive area of the tube.

Clearly though there are things with the breadboard version that will need changing for a complete, enclosed project.  One thing to change is the high voltage capacitors - the ones used on the mock-up are about 4x the physical size of the ones i've recently bought, yet have identical ratings. The neon lamp as part of the regulator does take up more space than zener diodes, but I will keep that - the glow is a perfect indicator that the circuit is live!


Thursday, 5 December 2019

More electronics from behind the curtain

I'm old enough to remember the fall of the Berlin wall (though not as many would have it the abandonment of Hadrians!) and it's because of that moment in history that today I could take delivery of a Soviet era SI-19BG Geiger tube.

This tube really is tiny! I would never have thought that a G-M tube could come through the post in a normal envelope! And that's wrapped up in bubble wrap!

I've done some tests on the breadboarded HV circuit now I have the 1Gohm resistor. The 3V circuit has trouble cutting back the current if final output regulation is used, but seems to happily give around 400V with the regulator chain connected at the first multiplier, dropping the current from 4.5mA to about 1.5mA.

I'm sure I can get this lower, with a bit of circuit tweaking. Right now though, the daft 11M impedance of my DMM means that the voltage reading on the LCD is not quite accurate,  so I hope to work out a way you correct this using parallel resistance to get a true 100:1 divider chain. I might check the impedance of some of my older less useful DMMs - I might be able to make a dedicated HV probe meter.

Edit - Well Doh! It isn't a parallel resistance I need! Its more series resistance of course! 1089MΩ in fact.  With the 1GΩ resistor this means I can make up the extra using standard low cost, low voltage resistors, plus a preset to allow precise calibration.

Monday, 2 December 2019

STS-6 tube arrives

The huge STS-6 tube has arrived safely from Volgograd, Russia- yet I'm still waiting for the delivery from RS Components!

I had hoped to have had at least one Geiger circuit working by now, but the late delivery of the 1GΩ resistor, plus some short notice IT work, have delayed me.